How we work

What we do before we list an add-on or publish a guide, what our tests do not cover, and who pays for this site.

How we choose add-ons

We start from what people use: official add-ons from the AI companies and other software makers, and popular open-source projects. Before we test one, we check three things:

Then we pick the few that matter most for each kind of work. Those are the must-haves, ranked from most useful to least.

What our test does

We install every add-on on a clean test system: an empty computer in a container, with no accounts, no keys and no access to our own files. We run the install steps from the project's own instructions and check that it starts. Each add-on page says exactly what we ran and what happened, with the date.

What our test does not cover

How we check guides and news

Every guide is written from the official documentation of the assistant. On the day we write or update a guide, we open each official page again. The guide shows that date and lists the official pages it is based on.

News items in Updates are checked against the primary source, such as the release notes or the company's own announcement, before we write them in our own words.

Who pays for this site

See all tested add-ons